Reproduce “Multi-Rate Flow Matching for Point Cloud Upsampling” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jingkai Shang, Edgar Torres, Duy Nguyen
MRFM upsamples sparse point clouds by assigning inference-time integration budgets to local patches while keeping the learned flow model fixed. A matched-budget permutation preserves step counts but worsens reconstruction when budgets are shuffled across patches. On PU1K at 4×, MRFM reports CD of 4.0434 ± 0.0213 versus 4.1887 ± 0.0209 for PUFM++; this demonstrates reconstruction quality, not faster inference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.